• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

On Intel Raptor Lake, any truth to the rumors that disabling all e-cores hurts single threaded performance of the p-cores??

you should post results with and without to prove it ? with temps, power use, averages and lows etc etc

from what I have seen from few that posted on youtube hardly any difference overall and they used windows 10 which isnt even optimized for e cores
 
Last edited:
Wasn't the thread stuff in W11 included in W10 in a fairly recent update, 12th Gen has been out for ages now so what might have been true at launch is no longer the case.

PC World YouTube done a recent test and found no difference between latest W11 and W10 in anything productivity, content creation or gaming but that was at stock, they didn't mess about with uncore etc but that was low single digit gains best case I think on 12th.
 
So I figured let's do some testing and see what we got. I did the following:

- Use my daily 13600k setup, ran YC2.5b to show it's actually running at those speeds and temps
- I then ran 3 runs of Cyberpunk 2077 with 1080p High but FSR manually disabled. I have a 4090 so no issues with having a GPU bottleneck
- Then went into bios and disabled all ecores
- I then looked for how much more Ring I could squeeze with everything else being the same. Just ecores disabled. That was 100mhz extra of ring.
- Then I ran the CP2077 bench 3 times again

Here's the data:
YC with ecores enabled:
image.png


YC with ecores disabled:
image.png


CP2077 with Ecores enabled (best of 3 runs):
image.png


CP2077 with Ecores DISABLED (best of 3 runs):
image.png



While the testing is limited, we can still deduce a few things:
- There isn't a major jump to be gained in ring performance with ecores disabled. 100mhz or even 200mhz is absolutely trivial
- There is no difference to the temps of the Pcores under a stressful load. In both scenarios, both peak around the same
- Clearly in CP2077, on a 13600k, disabling ecores is a BAD idea if performance is the goal
- This was done with a clean system with no background tasks and a fresh reboot. This is the best case scenario for the ecores disabled system

As always, do your testing instead of throwing out wild theories.
 
There is noting wrong with using ecores on Windows 10 and there is nothing wrong with using Windows 11 either, but whateva. You decided to hate on the hybrid for no reason whatsoever, so keep on hating
If someone is only running 8 threads, what is the point having additional cores drawing power?

Surely it would be better to redirect the power to the 8 P cores to run them faster?

Like in 2010/2011 the X5680 which had 6 cores, but 4 disabled so it could run at 4.4 GHz.
 
Last edited:
If someone is only running 8 threads, what is the point having additional cores drawing power?

Surely it would be better to redirect the power to the 8 P cores to run them faster?

Like in 2010/2011 the X5680 which had 6 cores, but 4 disabled so it could run at 4.4 GHz.
There is no such things as running 8 threads. Your computer wont be running only 8 threads.

If ecores are not being utilised then they wont draw any power.

Theories are worthless, testing is king. Ive tested on both my 12900k and the 13900k, disabling ecores is just silly. With the 12900k sure you could get like 2% better performnace in an extreme 480p scenario with ecores off cause of the ring frequency, but thats about it. With the 13900k you get absolutely nothing by disabling ecores.

I could only see a scenario that your ecores sp is extremely low while your pcores sp is extremely high, so you disable ecores in order to drop the voltage by a ton, but that's like what, 1 out a 1000 cpus.
 
There is no such things as running 8 threads. Your computer wont be running only 8 threads.

If ecores are not being utilised then they wont draw any power.

Theories are worthless, testing is king. Ive tested on both my 12900k and the 13900k, disabling ecores is just silly. With the 12900k sure you could get like 2% better performnace in an extreme 480p scenario with ecores off cause of the ring frequency, but thats about it. With the 13900k you get absolutely nothing by disabling ecores.

I could only see a scenario that your ecores sp is extremely low while your pcores sp is extremely high, so you disable ecores in order to drop the voltage by a ton, but that's like what, 1 out a 1000 cpus.
I meant 8 "cores", not 8 "threads". On Linux it's easy to limit no more than 8 cores are used.

And as the other guy said, we still want the full L3 cache. What was the consensus on this, do the remaining cores share the full L3?

So basically this new hybrid architecture lumps people with a performance cost they cannot avoid? If this wasn't true less cores would always guarantee higher performance?

I dont understand why so many E cores are needed? When we isolate cores in Linux we usually isolate ALL the OS tasks and HW interrupts to 1 or 2 cores. Certainly don't need 8/16 low priority cores! Did Intel do this because Windows 11 is so bloated?
 
Last edited:
I meant 8 "cores", not 8 "threads". On Linux it's easy to limit no more than 8 cores are used.

And as the other guy said, we still want the full L3 cache. What was the consensus on this, do the remaining cores share the full L3?

So basically this new hybrid architecture lumps people with a performance cost they cannot avoid? If this wasn't true less cores would always guarantee higher performance?

I dont understand why so many E cores are needed? When we isolate cores in Linux we usually isolate ALL the OS tasks and HW interrupts to 1 or 2 cores. Certainly don't need 8/16 low priority cores! Did Intel do this because Windows 11 is so bloated?
I dont get your question honestly. More cores are better for mt performance. If you don't care about mt performance you can get cpus with lower number of ecores, like the 13700k, the 13600k or the 12700k. Its like asking why does the 7950x has 16 cores, isnt 8 enough? Well sure for some people 8 are enough, so these people can buy the 7700x.

Why would less corea guarantee better performnace? Im not sure im following you
 
Last edited:
I dont get your question honestly. More cores are better for mt performance. If you don't care about mt performance you can get cpus with lower number of ecores, like the 13700k, the 13600k or the 12700k. Its like asking why does the 7950x has 16 cores, isnt 8 enough? Well sure for some people 8 are enough, so these people can buy the 7700x.

Why would less corea guarantee better performnace? Im not sure im following you

12700k has lower turbo frequency, 5.4GHz vs 5.8GHz

I'm after the highest frequency. Why must this come saddled with 16 low-priority cores? Surely it should be the opposite, higher turbo frequency comes with less E-cores?

Otherwise, someone like myself and the other guy in this thread are asking..... how fast would the i9 be without those 16 E cores? And how much are those (pointless) E cores costing?
 
Last edited:
Wasn't the thread stuff in W11 included in W10 in a fairly recent update, 12th Gen has been out for ages now so what might have been true at launch is no longer the case.

PC World YouTube done a recent test and found no difference between latest W11 and W10 in anything productivity, content creation or gaming but that was at stock, they didn't mess about with uncore etc but that was low single digit gains best case I think on 12th.


Any source on this?

Only WIN11 is thread director aware.

Certainly not WIN10 21H2.

Not sure about WIN10 22H2?
 
But how exactly do you have more headroom in tasks that use 6 cores? Lol that doesnt make any sense. If your task is using only 6 cores then ecores should be parked and basically doing nothing, consuming no wattage...

Have you actually tried it or are you talking hypothetically? Cause i have, makes absolutely no difference in games, please stop spreading missinformation.

Those tasks usually use 6 cores, but some may use 8 but no games use more than 8. And better to have a much faster P core waiting to be called on than e-waste cores around, And P cores ring clocks higher and they have more thernal headroom for a static manual overclock with e-waste cores disabled.
 

Seems pretty conclusive that there is no difference with win 10 to win 11 and you have armies of people in here saying leave e-cores on.


Well that is luck. WIndows 10 is not aware of the thread director, so potential exists for trouble with scheduling. Though if you use process Lasso and just run an app that straight up uses all threads it will not matter. The issue is more with overall scheduling of background tasks and such if you do set and forget without using Process Lasso.

And if you want more thermal headroom for P cores, e-cores off is a good thing especially if doing a static al P core speed all the time and care about stressing all core all the time work load for stability test load only like an old school overclocker.
 
Those tasks usually use 6 cores, but some may use 8 but no games use more than 8. And better to have a much faster P core waiting to be called on than e-waste cores around, And P cores ring clocks higher and they have more thernal headroom for a static manual overclock with e-waste cores disabled.
Prove it. Post a benchmark with ecores off, ill post the same with ecores on, lets see what happens.
 
I've seen games utilise all 4 e-cores on my 12700KF as well as use a bunch of p-cores - Obviously not the way it should be working but game engines haven't really been optimised properly for a long time now to maximumise multi-core CPUs anyway so this is the world we live in.

I've also seen the e-cores in other games steadily punch away with background apps like RTSS and other apps exactly as you would expect, and exactly as designed whilst the p-cores focus on the foreground app, the game.

Have yet to see anything concrete that shows disabling the e-cores results n better performance as an outright thing.
 
I've seen games utilise all 4 e-cores on my 12700KF as well as use a bunch of p-cores - Obviously not the way it should be working but game engines haven't really been optimised properly for a long time now to maximumise multi-core CPUs anyway so this is the world we live in.

I've also seen the e-cores in other games steadily punch away with background apps like RTSS and other apps exactly as you would expect, and exactly as designed whilst the p-cores focus on the foreground app, the game.

Have yet to see anything concrete that shows disabling the e-cores results n better performance as an outright thing.

WIN10 or WIN11 are you running?

WIN11 never put background tasks om e-cores and neither did WIN10 when I tried them. They are useless for that as such.

Though I have heard WIN11 is supposed to be able to do it. WIN10 has no clue how to.

The e-cores are there for assistance for perfectly parallel apps that take as many compute cores as you can throw at them. For gaming they do nothing and can hurt on WIN10 as they drag ring clock down.

Disabling them is wise on WIN10 unless you use Process Lasso and run lots of productivity apps that are perfectly parallel.

WIN11 not so much because of this: https://fox-laptop.com/pc-component...es-but-without-them-everything-is-only-worse/
 
WIN10 or WIN11 are you running?

WIN11 never put background tasks om e-cores and neither did WIN10 when I tried them. They are useless for that as such.

Though I have heard WIN11 is supposed to be able to do it. WIN10 has no clue how to.

The e-cores are there for assistance for perfectly parallel apps that take as many compute cores as you can throw at them. For gaming they do nothing and can hurt on WIN10 as they drag ring clock down.

Disabling them is wise on WIN10 unless you use Process Lasso and run lots of productivity apps that are perfectly parallel.

WIN11 not so much because of this: https://fox-laptop.com/pc-component...es-but-without-them-everything-is-only-worse/
From your freaking link, with ecores off the 12700k was slower in every single benchmark bar one.....including games. So wtf are you talking about ???
 
WIN10 or WIN11 are you running?

WIN11 never put background tasks om e-cores and neither did WIN10 when I tried them. They are useless for that as such.

Though I have heard WIN11 is supposed to be able to do it. WIN10 has no clue how to.

The e-cores are there for assistance for perfectly parallel apps that take as many compute cores as you can throw at them. For gaming they do nothing and can hurt on WIN10 as they drag ring clock down.

Disabling them is wise on WIN10 unless you use Process Lasso and run lots of productivity apps that are perfectly parallel.

WIN11 not so much because of this: https://fox-laptop.com/pc-component...es-but-without-them-everything-is-only-worse/
Windows 11, and your post makes no sense because none of that is observed behaviour. in actual practice. Thread director works perfectly and background tasks offload to e-cores effectively, exactly as intended.
 
Back
Top Bottom